Periodontitis is a gum disease that is the result of infections and inflammation of the gums and the bone which surround and support the teeth. In the most severe cases, teeth may loosen or fall out. Using the ensemble machine learning method Random Forest we predict the risk of developing Periodontitis with three different models and also investigate which predictors are the most useful for the predictions. Each model is based on the same data with both systemic and local predictors but the predictors included in each model differ. The model's performances are evaluated based on different types of accuracy measures and the predictors are evaluated using Variable Importance plots and measures. The model with the most predictors included performed best, as a result of it using the most local predictors which seemed to have a higher importance in predicting, compared to the systemic predictors.
| Identifer | oai:union.ndltd.org:UPSALLA1/oai:DiVA.org:uu-503326 |
| Date | January 2023 |
| Creators | Waldenfjord, Noel, Sedwall, Albert |
| Publisher | Uppsala universitet, Statistiska institutionen |
| Source Sets | DiVA Archive at Upsalla University |
| Language | English |
| Detected Language | English |
| Type | Student thesis, info:eu-repo/semantics/bachelorThesis, text |
| Format | application/pdf |
| Rights | info:eu-repo/semantics/openAccess |
Page generated in 0.123 seconds